DataMeadow DataMeadow A Visual Canvas for Analysis of Large-Scale Multivariate Data Niklas Elmqvist – John Stasko – Philippas Tsigas IEEE Symposium on Visual Analytics Science & Technology 2007

November 1, 2007 DataMeadow: Elmqvist, Stasko and Tsigas2 In media res…  Let’s start with a demo!

November 1, 2007 DataMeadow: Elmqvist, Stasko and Tsigas3 Parallel Coordinates  First proposed by Alfred Inselberg in The Visual Computer in 1985  Basic idea: stack dimension axes in parallel, points become polylines  Advantage: easy to add new dimensions  As we add dimensions, the parallel coordinate diagram grows horizontally  Another solution is to transform to polar coordinates and make radial axes  Starplot diagram

November 1, 2007 DataMeadow: Elmqvist, Stasko and Tsigas4 Elements and Dependencies  Visual Elements  Conforms to system-wide data format  Visual appearance depending on input  Input and output ports  Three types: Sources, sinks, transformers  Dependencies  Conforms to system-wide data format  Directed link between two elements  Propagates data from source to destination  Interactive updates ? … … … …

November 1, 2007 DataMeadow: Elmqvist, Stasko and Tsigas5 The DataRose  2D starplot display  Transformer visual element  Average as black polyline  Shows data distribution using different representations  Opacity bands [Fua et al. 1999]  Color histogram bands  Parallel coordinates

November 1, 2007 DataMeadow: Elmqvist, Stasko and Tsigas6 DataRose Representations  Parallel coordinate mode  See all details  Cannot see distribution  Color histogram mode  LOCS color scale  Brightness = high value

November 1, 2007 DataMeadow: Elmqvist, Stasko and Tsigas7 DataRose Types  DataRoses can be of several different types  Each type represents a specific multi-set operation  Four types:  Source: external database loaded from a file  Union: all input cases combined  Intersection: input cases that exist in all input sets  Uniqueness: input cases that exist in one input set

November 1, 2007 DataMeadow: Elmqvist, Stasko and Tsigas8 Viewers and Annotations  Viewers are sink elements: accept input - no output  Shows quantitative information for the inputs  Barchart  Piecharts  Histogram  Annotations support communication of analyses  Labels  Notes  Images  Reports

November 1, 2007 DataMeadow: Elmqvist, Stasko and Tsigas9 Evaluation  Evaluation: expert review (think-aloud protocol)  Participants: two visualization researchers  Dataset: US Census 2000  Three types of open-ended questions:  Direct facts: “What is the average house value in Georgia?”  Comprehension: “Which state has the highest ratio of small and expensive houses?”  Extrapolation: “Is there a relation between fuel type and building size in Alaska?”  Results: positive, has lead to new design iterations  Participants were able to solve questions  Interaction quoted as main benefit

November 1, 2007 DataMeadow: Elmqvist, Stasko and Tsigas10 Contributions  A highly interactive visual canvas (DataMeadow) for multivariate data analysis using multiple small visualization components  A visual representation (DataRose) based on axis- filtered parallel coordinate starplots that can be linked together into interactive visual queries  Results from a qualitative user study showing the use of our system for multivariate data analysis

November 1, 2007 DataMeadow: Elmqvist, Stasko and Tsigas11 Future Work  Additional visual components for the DataMeadow  More complex visual queries  Additional annotation and communication support  Non-standard input devices  Pen-based interfaces  Non-standard output devices  Large displays  Collaborative visual analytics

November 1, 2007 DataMeadow: Elmqvist, Stasko and Tsigas12 Questions?  Contact information: Niklas Elmqvist WWW: Phone: Fax: Pictures courtesy of Helene Gregerström Taken at the Atlanta Botanical Gardens